Study On Parameter Identification And SOC Estimation Of Ni/MH Battery For EV

Since energy and environmental issues become increasingly serious, the electric vehicle (EV) with the advantages of zero emissions and low noise has attracted a great deal of attention over the world. Battery energy management system (BMS) is one of the key technologies which impedes the commercialization and practical application of EV, while the prediction of state of charge (SOC) is the main task of BMS. The exact and reliable estimation of SOC of the battery is necessary and important for BMS to run well in EV. Based on the study of battery model and parameter identification, we are committed to the SOC estimation of the nickel metal hydride (Ni/MH) for EV.Firstly, the background of the project is briefly introduced. Then by comparing the current various power storage batteries against many performance indexes, the Ni/MH battery is found to be a quite ideal power source for electric vehicles. After this, the definition of the SOC is given, and several different kinds of SOC estimation methods are introduced. In addition, the key content of this paper is summarized.Secondly, the history and principle of the Ni/MH battery are briefed. The characteristics of the battery, such as the battery voltage, the resistance, the capacity are analyzed with the experimental curves. In the introduction of the capacity, the main factors which impact the SOC and the corresponding measures are pointed out.Building a good battery model is an important way to improve the accuracy of the SOC estimation. So some common models of batteries are analyzed and the second-order RC model is selected according to the battery's response against pulse current. Simulation results show that the second-order RC model has better dynamic characteristics. In order to improve the accuracy of the model, the values of the model parameters are fitted out using exponential function, and the real time parameter identification is implemented using the methods of finite memory recursive least squares and kalman filter respectively.Finally, based on the current integration model and the second-order RC model, the state-space model is educed. After this, SOC is estimated by the extended kalman filter (EKF). The results of simulation are given, and the disturb-resisting capability and robustness of the algorithm are tested. At the end, parameter and state estimation together is accomplished, that is parameter identification is carried out first and the parameters are used to estimate SOC later. The results make it clear that this method has better estimation accuracy than EKF.